A startup called Near Space Labs is turning stratospheric balloons into flying robots that use AI to speed up insurance claims after disasters.

Why it matters: Hurricane season in the U.S. is here and many are bracing for the complicated aftermath of collecting insurance for property loss as well as the physical impact of potential disasters.

How it works: Near Space builds and deploys autonomous devices with proprietary cameras and sensors that capture all kinds of ultra-precise data from disaster areas.

• The devices, known as Swift robots, travel in the stratosphere — higher than airplanes — at a vantage point that makes it easy to map large areas more cost-effectively than planes or drones.
• The balloons can launch from afar without entering unsafe areas.
• Near Space says insurers use the data collected by the robots to price risk accurately, allowing communities to stay insurable and get paid faster.

Between the lines: The climate crisis has spurred an insurance crisis with soaring premiums and carriers retreating from high-risk states like California and Florida.

• Homeowners lose access to coverage or face claim denials based on outdated or low-res imagery. Near Space data can show clear images of new roofs or other upgrades owners have made to prepare for disasters.

By the numbers: Near Space was founded in Brooklyn in 2017 and has raised over $40 million in funding, including a $20 million Series B last month.

• It would take more than 800,000 drones to capture what one Swift robot can capture, Near Space co-founder and CEO Rema Matevosyan told Axios.
• Matevosyan says Swift robots can be mounted on balloons in roughly 10 minutes. The rest of the process is autonomous.
• Near Space provides coverage of 80% of the continental United States population.

Context: Several startups are using AI and advanced imaging to streamline the labyrinthine insurance industry.

• ZestyAI, an AI-powered property and climate risk analytics company that launched in 2021, works with California's FAIR Plan, the insurer of last resort.
• Arturo uses aerial and satellite imagery combined with AI to help insurers reduce costs.
• While not directly involved in the insurance industry, Sceye uses stratospheric balloons to track extreme weather and climate disasters such as wildfires.

What's next: The company is working on genAI agents designed to allow insurers to query Near Space Labs' imagery.

• For example, you could potentially ask an agent to detect all of the tarps put up on roofs after a hurricane — a simple and quick way to help large insurers understand exposure to a disaster.

Yes, but: Better data helps with pricing and claims, but it doesn't stop insurers from pulling out of high-risk markets entirely.

• If climate risk becomes truly uninsurable, even perfect imagery won't keep homes covered.
• High-res images of homes also pose privacy risks. Near Space assured Axios that its technology cannot see through walls or inside buildings, following the same privacy standards used by other aerial photography companies for decades.
• "Near Space protects its imagery with comprehensive data governance frameworks and carefully selects partners who adhere to documented privacy and security protocols," the company said in an email to Axios.

What they're saying: Matevosyan says her goal is to use tech to provide equity and access, especially in remote, underserved communities, often missed by satellite or aerial surveillance.

• "There's no climate resilience without financial resilience," Matevosyan says. "Financial resilience is insurance in the face of catastrophes."
• "Frequent data helps us understand the planet and helps us understand this ever-evolving risk. We can actually apply methods that we know work and help everybody sort of adjust to this new reality, which is that our environment is going to continue changing."

Go deeper: How climate change screws up mortgage markets

David Sacks, the venture capitalist advising President Trump on crypto and AI policy, is divesting his holdings in foundational AI companies and hyperscalers like xAI and Meta, according to a White House memo released Friday.

• The White House previously disclosed that Sacks and his firm, Craft Ventures, divested around $200 million of crypto-related assets.

Why it matters: Sacks entered the Trump administration with a slew of investments that could have created conflicts of interest, or at least perceptions of them, but has taken steps to minimize that risk.

Zoom in: The memo appears to have been written months ago, since it refers to March 31 as a future date.

• It says that Sacks was required to sell stock in such publicly traded companies as Amazon and Taiwan Semiconductor Manufacturing by the end of Q1, and was set to divest his Meta shares by the end of this month (at the latest).
• Craft Ventures sold stakes in 10 AI companies. This includes xAI — which is listed as a "pending" divestiture in the memo, but a firm spokesperson says the deal is done.
• He and Craft are retaining positions in dozens of SaaS and hardware startups, some of which either already use AI or are expected to.

The bottom line: Both this memo and the earlier one focused on crypto are waivers that let Sacks continue working as a special government employee.

During our recent interview, Anthropic CEO Dario Amodei said something arresting that we just can't shake: Everyone assumes AI optimists and doomers are simply exaggerating. But no one asks:

• "Well, what if they're right?"

Why it matters: We wanted to apply this question to what seems like the most outlandish AI claim — that in coming years, large language models could exceed human intelligence and operate beyond our control, threatening human existence.

That probably strikes you as science-fiction hype.

• But Axios research shows at least 10 people have quit the biggest AI companies over grave concerns about the technology's power, including its potential to wipe away humanity. If it were one or two people, the cases would be easy to dismiss as nutty outliers. But several top execs at several top companies, all with similar warnings? Seems worth wondering: Well, what if they're right?
• And get this: Even more people who are AI enthusiasts or optimists argue the same thing. They, too, see a technology starting to think like humans, and imagine models a few years from now starting to act like us — or beyond us. Elon Musk has put the risk as high as 20% that AI could destroy the world. Well, what if he's right?

Keep reading.

• Exclusive: OpenAI execs will convene privately in D.C. today with titans from oil, energy and finance to strategize on AI's increasing power needs. (Axios)
• Google and other big Scale AI customers plan to cut back their spending with the company in the wake of Meta's giant investment in it. (Reuters)
• Kashmir Hill has a must-read on how ChatGPT led some people down conspiratorial rabbit holes. (NYT)
• Meta CTO Andrew Bosworth and OpenAI product chief Kevin Weil are among the tech executives who have signed up for Detachment 201, an Army Reserve unit aimed at helping the military get more tech-savvy. (WSJ)
• TikTok's new AI ad tools will let marketers generate video ads for the service with only a text prompt and a product photo. (Bloomberg)

A troop of second grade Girl Scouts in New Jersey created a spot for people to grab or drop off reusable bags near a Trader Joe's, along the lines of the Little Free Library bookcases.